Sunda Java Trench Kinematics
-
Upload
maximillian-heartwood -
Category
Documents
-
view
232 -
download
0
Transcript of Sunda Java Trench Kinematics
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 1/13
Sunda-Java trench kinematics, slab window formation and overriding
plate deformation since the Cretaceous
J.M. Whittaker ⁎, R.D. Müller, M. Sdrolias, C. Heine
EarthByte Group, School of Geosciences, University of Sydney, Sydney, Australia
Received 20 June 2006; received in revised form 12 December 2006; accepted 13 December 2006
Editor: R.D. van der Hilst
Available online 19 January 2007
Abstract
The kinematics and time-dependence of back-arc extension or compression is one of the most poorly understood aspects of
plate tectonics, and has nearly exclusively been studied from snapshots of present-day observations. Here we combine absolute and
relative plate motions with reconstructions of now subducted ocean floor to analyse subduction kinematics and upper plate strain
from geological observations since 80 Ma along the 3200 km long Sunda-Java trench, one of the largest subduction systems on
Earth. Combining plate motions and slab geometries enables us to reconstruct a time-dependent slab window beneath Sundaland,
formed through Wharton spreading ridge subduction. We find that upper plate advance and retreat is the main influence on upper
plate strain, but subduction of large bathymetric ridges, and slab-window effects, also play a significant, and at times dominant,
role. Compression in the Sundaland back-arc region can be linked to advance of the upper plate. Extension of the Sundaland back-
arc region correlates with two patterns of upper plate motion, (a) retreat of the upper plate, and (b) advance of the upper plate
combined with more rapid advance of the Sundaland margin due to hinge rollback. Subduction of large bathymetric ridges causes
compression in the upper plate, especially Wharton Ridge subduction underneath Sumatra over the period 15–0 Ma. Our
reconstructions unravel the evolving geometry of a slab window underlying the Java–South Sumatra region, and we propose that
decreased mantle wedge viscosities associated with this slab-window exacerbated Palaeogene extension in the Java Sea region via
active rifting, and enabled Sumatran continental extension to continue at 50–35 Ma when upper plate advance would otherwise
have led to compression.
© 2007 Published by Elsevier B.V.
Keywords: Sundaland; subduction; kinematics; slab window
1. Introduction
Numerous models have been proposed to account for
the time-dependence of extension and back-arc basin
formation in a subduction setting, including: (1) the
extrusion model, where back-arc formation is related to
lateral absolute upper plate motion [1–3] (2) the sea-
anchor model, where back-arc formation is related to the
force generated by the down-going slab resisting lateral
motion [2], (3) magmatic models, where back-arc
formation is related to mantle flow in the wedge
overlying the slab [4–6], and (4) the slab pull model,
where back-arc formation is related to subduction hinge
rollback caused by negative buoyancy of the subducting
slab [7–9]. The “sea-anchor ” model and the “extrusion”
Earth and Planetary Science Letters 255 (2007) 445–457
www.elsevier.com/locate/epsl
⁎ Corresponding author.
E-mail address: [email protected] (J.M. Whittaker).
0012-821X/$ - see front matter © 2007 Published by Elsevier B.V.doi:10.1016/j.epsl.2006.12.031
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 2/13
model are the more likely models for back-arc formation
according to analysis of recent mantle flow data from
the Mariana subduction system [10]. Strain in the “ back-
arc” region of an upper plate in a subduction setting can
range from strongly compressional to seafloor spread-
ing. In order to examine relationships between various plate kinematics parameters and a range of upper plate
strain regimes, we focus on the Sunda-Java subduction
zone through the Cenozoic. Excluding Andaman Sea
back-arc spreading from∼13 Ma to the present [11], the
Sundaland margin has experienced compression, non-
rift related subsidence, or crustal extension during the
Cenozoic [12].
In both the “sea-anchor ” model and the “extrusion”
model a major parameter that influences upper plate
strain is upper plate motion. Upper plate motions have
long been associated with upper plate strain [1], withextensional back-arcs correlating with retreating upper
plates, and compressional back-arcs correlating with
advancing upper plates e.g. [2,13–15]. We have
compared reconstructed absolute and trench-normal
upper plate motions with strain regimes known to
occur in the Sundaland margin to 60 Ma.
Recently, Lallemand and Heuret [13] observed a
correlation between present-day shallow slab dip angles
(0–125 km) and upper plate strain, with compressive and
extensional regimes correlating to low and high shallow
slab dips, respectively. Sdrolias and Müller [14] found
that a combination of factors, including shallow slab dip(0–100 km), influences the initiation of back-arc
spreading and that once initiated, back-arc spreading
continues regardless of upper plate motion. Sdrolias and
Müller [14], also observed a correlation between age of
present-day, non-perturbed subducting lithosphere and
shallow slab dip angle (0–100 km), where older subduc-
ting lithosphere correlates with steeper shallow slab dip
angles and vice versa. We have utilised this relationship to
reconstruct shallow slab dips for the Sunda-Java trench
back to 80 Ma, using palaeo-age grids. We then compare
the reconstructed shallow slab dips with mapped strainregimes through time along the Sundaland margin.
Complicating kinematics at the Sunda-Java trench
has been the subduction of the active Wharton Ridge
from ca. 70 Ma [16], to 43 Ma [17], and the remainder of
the then extinct Wharton Ridge representing a bathy-
metric ridge from 43 Ma to the present. During
subduction of an active ridge, a slab window may
form under the upper plate [18,19]. A slab window
develops when down-going plates continue diverging
but trailing plate edges cease to grow and may even
become hot and begin to melt [19]. The slab window
widens as this process continues. Slab windows
occurring beneath the west coast of North America
(e.g. [20–24]), Central America (e.g. [25]) and South
America (e.g. [26,27]) have been well studied compared
to the Indonesian subduction zone. The palaeo-positions
of slab window are normally estimated using geological
data from the overriding plate, such as changes involcanism and tectonic events such as regional uplift.
For the Sunda-Java trench we have used time-dependent
plate motion vectors combined with reconstructed
palaeo-age grids to compute the size and position of
the slab window beneath the southern Sundaland
margin.
2. Methods
In our study, we split the Sundaland margin into three
regions; the Andaman Sea, Sumatra, and Java (seeFig. 1). For each region, we identified and summarized
different periods of upper plate strain from Morley [28],
Bishop [29], Hall [30], Letouzey et al. [12], Eguchi et al.
[31], and Curray et al. [11] (see Fig. 3) and categorised
each period based on the method of Jarrard [15].
We calculated shallow slab dip angles at each point
along the Sundaland trench for 5 million year time
stages from the present to 80 Ma, utilising the
relationship y =0.1961 x + 12.232, where x is age of
subducting lithosphere at the trench, and y is the shallow
slab dip angle [14]. We obtained the age of the down-
going lithosphere at the trench from revised versions of oceanic palaeo-age grids from Heine et al. [16]. From
the calculated shallow slab dip angles, average shallow
slab dip angles were computed for each region at all time
stages. Averaging the slab dip angles minimizes any
distortions in slab dip angle due to proximity to the edge
of the subducting plate, which can be around 10° steeper
close to slab edges [13].
For each stage we calculated absolute plate motion
vectors for four plates; the Sundaland plate (80–0 Ma), the
Sumatra-Java plate (or Sundaland margin) (80–0 Ma), the
Indian plate (80–45 Ma), and the Australian plate (80–0 Ma) (see Fig. 1). Vectors were calculated at points every
500 km along the Sundaland trench, which was digitised
for the present day then fixed to the Sumatra-Java plate for
rotation back through time. Our plate reconstructions used
the Heine et al. [16] and Gaina and Müller [32] plate
kinematic models in a moving hotspot reference frame
from O'Neill et al. [33]. Absolute plate motion vectors
were used to calculate trench-normal vectors for each
plate, which are useful for examining the overall compres-
sional/extensional forces acting on the continental margin.
We reconstructed the position of the slab window
beneath Sundaland, using the method of Thorkelson
446 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 3/13
[18] which calculates the slab-window position using
ridge-transform geometry and convergence vectors. Our
absolute plate motion vectors were used to calculate
convergence vectors, which were used in conjunction
with the reconstructed location of the Wharton Ridge
through time from the palaeo-age grids to establish the
shape and size of the slab window. Vectors used tocalculate the shape of the slab window are presented in
velocity–space diagrams in Fig. 2 for time slices where
there was an active triple junction at the Sundaland
margin. Generally, the geometry of a slab window is
affected by relative plate motions, pre-subduction ridge-
transform fault geometry, subduction angles, thermal
erosion, deformation caused by spherical shell stress and
lateral and down-dip changes in the angle of slab dip
[18]. In our approach, we have assumed, a horizontal
subducted slab, no thermal erosion of the diverging plate
edges, and no deformation from spherical shell stress.
Due to these assumptions we have calculated a
minimum slab window, as adding the effects of a
dipping slab and thermal erosion of plate edges would
likely lead to a larger extent for the slab window. We
have also limited the lateral extent of the slab window to
1000 km perpendicular to the trench because at this
point the slab can be assumed to have reached the
660 km mantle discontinuity and the slab windowwould no longer have a discernable affect on the
overriding plate.
3. Plate kinematics and overriding plate deformation
3.1. Plate motions
Fig. 2 illustrates that from 80 Ma to the cessation of
Wharton Ridge spreading (∼43 Ma [17]), the Australian
Plate moved at a much slower rate than the Indian Plate
(Fig. 2(i–vii)). Initial plate and margin geometry and
kinematics of the India–Eurasia collision remains
Fig. 1. Topographic and bathymetric map of SE Asia, SF — Sumatra Fault, boundaries of the Sundaland margin (dashed) and the Sundaland core
(solid), small arrows depict motion of Sumatran fore-arc northwest and southeast of Batu Island from Prawirodirdjo et al. [48].
447 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 4/13
controversial [34], but it is generally accepted that an
India–Eurasia related collision slowed northward Indian
Plate motion from 60–55 Ma (e.g. [34–36]). Fig. 2(i–viii)
shows the Indian Plate moving rapidly from 80 Ma until
absolute plate motions dropped from an average of
120 mm/yr at 60–65 Ma to 81 mm/yr by 45–50 and26 mm/yr by 40–45 Ma. Indian plate motion increases
from the low at 40–45 Ma to an average of 69 mm/yr at
30–35 Ma decreasing to 63 mm/yr at 15–20 Ma.
An advancing upper plate is expected to cause com-
pression in the upper plate margin. Our reconstructions
show two different types of advancing upper plate motion.
The first type, where both the upper margin and upper
core advance at the same rate, we call ‘uniform upper plate
motion’. Uniform upper plate motion occurs for Java from
15 Ma to the present (Fig. 3(viii–x)), which corresponds
with compression that is known to have affected Javafrom∼15 Ma to the present day. This type of upper plate
advance also occurs at all points along the Sunda-Java
trench from 80 Ma to ∼60 Ma, so it is possible that the
entire Sunda-Java margin was affected by compression
during the period 80–60 Ma. However, the presence of an
underlying slab window due to the subduction of the then
active Wharton Ridge from 70 Ma may have had an effect
on the southeastern portion of the margin.
The second, where reconstructions show the upper
margin advancing more rapidly than the upper core, we call
‘differential upper plate motion’. Differential upper plate
motion occurs for the Andaman Sea at 30–15 Ma (Fig. 3(i–iii), Sumatra, strongly at 30–15 Ma (Fig. 3(iv–vii)),
southern Sumatra, weakly at 50–30 Ma (Fig. 3(vi–vii)),
and Java at 45–15Ma(Fig. 3(viii–x)). The upper margin is
known to have experienced extension during all of these
periods, which is expected to occur on the upper plate as
the margin draws away from the core. Subduction hinge
rollback, where the retreating hinge draws the margin away
from the core, is the most likely explanation for this pattern
of upper plate behaviour. Subduction hinge-roll has pre-
viously been suggested as the major mechanism causing
Malay–Thai basin extension at 30–15 Ma [28]. Non-rift
related subsidence, observed in Java and Sumatra from
15–20 Ma can be explained as the period of change over
from extension to compression.
The increase in rate of advance of the Sundaland core
at 30–20 Ma at almost all reconstructed locations shown
in Fig. 2 is likely to be a product of extrusion caused byIndia–Eurasia collision. Southeasterly extrusion of
Sundaland from 40 Ma caused by the India–Eurasia
collision are supported by mantle tomography images
showing a change in present-day slab structure at 700
and 1100 km depth [37,38]. ‘Hard’ collision between
India and Eurasia is thought to have caused the slow-
down in northward Indian motion at 30–20 Ma [39–41].
Morley [28] shows that more southerly blocks of
Southeast Asia were “squeezed out faster than their
more northerly neighbours”. This pattern fits increased
rates of upper plate advance observed at ∼
50 Ma and30–20 Ma (Fig. 3), which subside following the initial
extrusion related increase.
Our reconstructions (Fig. 3(i–iii)) show retreat of the
Andaman Sea upper plate at 15–0 Ma. This correlates
with Andaman Sea back-arc spreading that has occurred
from ∼13 Ma [11]. This correlation has previously been
observed by Sdrolias and Müller [14], who also found
that once initiated back-arc spreading is not affected by
motion of the upper plate and noted that it is believed
that spreading in this location is controlled by India–
Eurasia related extrusion tectonics.
Figs. 2(v–vi) and 3 show a reversal in Sundaland coremotion, from retreating at 55 Ma to advancing by
∼47 Ma. This change in motion is a consequence of
extrusion tectonics caused by the India collision. Upper
plate retreat occurs at all points along the Sunda-Java
trench except the northern Andaman Sea at 60–50 Ma
(Fig. 3). In the Java region this corresponds with a
known period of extension. Retreating upper plates have
long been associated with extension in back-arc areas
[2,13–15]. To initiate, back-arc spreading requires not
only a retreating upper plate but age of down-going
lithosphere at the trench N55 Myr, and a shallow slab
Fig. 2. Reconstructed absolute plate motions of the overlying Sundaland plate, and the down-going Indo–Australian plate from 80 Ma to the present.
Arrows represent stage (5 Myr) motions of the Sundaland margin (blue arrows), and Indian (Australian) plate (red arrows), and correspond to; (i) 80 –
75 Ma (ii) 75–70 Ma (iii) 70–65 Ma (iv) 65–60 Ma (v) 60–55 Ma (vi) 55–50 Ma (vii) 50–45 Ma (viii) 45–40 Ma (ix) 40–35 Ma (x) 35–30 Ma (xi)
30–25 Ma (xii) 25–20 Ma (xiii) 20–15 Ma (xiv) 15–10 Ma (xv) 10–5 Ma, and (xvi) 5–0 Ma. Tectonic regimes shown by compressional (black
centred ‘ beachball’), extensional (white centred ‘ beachball’) and subsidence (denoted by ‘sub.’) symbols (symbols are not oriented, symbolic only)
for the Java, south Sumatra and Andaman Sea regions where information was available from the literature [12,29–31]. The location of the mid-ocean
ridge and velocity–space diagrams (upper right of each figure) are shown for time slices when there was an active triple junction at the Sundaland
Trench, and hence a growing slab window. In the velocity–space diagrams; I — Indian Plate; R — Ridge; S — Sundaland Plate; A — Australian
Plate; H — hotspot; J — previous ridge–trench intersection location. The reconstructed positions of the slab window due to the subduction of the
Wharton Ridge (active until 43 Ma) are shown by thick black lines, dashed sections where slab window cut off at distance 1000 km from the trench.Thin black line represents our Sunda-Java trench location.
448 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 5/13
dip N30° [14]. Fig. 2(v–vi) shows that from 60–50 Ma,
the age of subducting lithosphere at the trench was
b55 Myr at all points on the trench southeast of point
1500, which is insufficient to initiate back-arc spreading
at this time. Therefore, we predict that crustal extension
affected the southeastern Sundaland back-arc east of
point 1500 for the period 60–50 Ma. For the southern
Andaman Sea section of the trench (points 1000–1500
449 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 6/13
Fig. 2(v–vi)) the age of subducting lithosphere at the
trench was N55 Myr, and Fig. 3(ii–iii) shows upper
plate retreat. However, our reconstructed slab dip for the
Andaman Sea for 60–50 Ma is ∼22°, suggesting that
conditions were not conducive to back-arc spreading
initiation in the Andaman region at this time.
Central and northern Sumatran upper plate retreat
suggests that the Sumatra back-arc should have
Fig. 2 (continued ).
450 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 7/13
experienced extension from 15 Ma to the present (Fig. 3
(iv–vii)). However, compression is known to have
affected this area during this period. It is likely that the
presence of the subducti ng Wha rton Ridge and
Investigator Fracture Zone (IFZ) are responsible for
this deviation from the expected upper plate regime.
Fig. 2 (continued ).
451 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 8/13
3.2. Bathymetric ridge subduction
The Wharton Ridge first subducts beneath eastern Java
[16], at 70 Ma, which likely caused the Sundaland margin
to rotate clockwise about a rotation pole close to the area
at this time. Presently, the Wharton Ridge and Investigator
Fracture Zone IFZ subduct beneath northern–central
Sumatra (Fig. 1). The subduction of bathymetric features
Fig. 2 (continued ).
452 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 9/13
Fig. 3. Trench-normal component of the reconstructed absolute plate velocity plotted for the Sundaland core and margin for the period 0–80 Ma at
points every 500 km along the Sunda-Java trench (see Fig. 2). Positive values indicate advance (oceanward motion) and negative values indicate
retreat (landward motion) of the upper plates. Background shading represents known tectonic regimes for the Sundaland back-arc summarised from
Morley [28], Bishop [29], Hall [30], Letouzey et al. [12], Eguchi et al. [31], and Curray et al. [11], where (1) compression is represented by crosses,
(2) subsidence is represented by dots, (3) extension is represented by light shading, and (4) spreading is represented by dark shading.
453 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 10/13
is widely accepted to cause broadly distributed deforma-
tion in the fore-arc [42,43]. Geodetic strain and rotation
rates show that the northern Sumatran region currently
endures a highly compressive regime [44]. Rates of
seismic deformation are at a maximum offshore Sumatra
near Nias Island (5.2± 0.65 mm/yr) and progressivelydecrease northward (1.12 ± 0.13 mm/yr) [45]. Due to
oblique subduction and extension to the north, Sumatra,
and the Sumatran fore-arc, are divided into a series of
NW–SE striking slices that move towards the northwest,
separated by right-lateral faults [46]. Most displacement
on these faults occurs in northwest Sumatra and dissipates
towards the southeast [47]. Geodetic observations from
GPS data [48] reveal an interesting change in Sumatran
fore-arc motion centred around Batu Island (Fig. 1).
Southeast of Batu Island, the Sumatra fore-arc moves
northeast, roughly parallel with the motion of the Indian plate, while northwest of Batu Island the Sumatran fore-
arc moves to the northwest [48]. This change in fore-arc
motion has been ascribed to decoupling between the
northern fore-arc and mantle wedge due to increased pore
pressures in the fore-arc thrust fault due to subduction of
thick Nicobar fan sediments [48].
The Wharton Ridge subducts beneath Nias Island
where seismic deformation is highest and the IFZ
subducts directly beneath Batu Island where the Sumatran
fore-arc begins to move in a northwest direction. Thus,
subduction of the Wharton Ridge and IFZ is another
mechanism causing the high seismic deformation rates,change in fore-arc motion, and concentration of strike-slip
motion that occurs in northern Sumatra. Fig. 3(iv–vii)
shows that rate of Sumatran upper plate retreat at 15–0 Ma
is not rapid (0–5 mm/yr), so it is likely that extension
experienced by the Sumatra back-arc from this mecha-
nism is relatively small. It is possible that present-day
compression from subduction of the Wharton Ridge and
Investigator Fracture Zone dominates over extension
resulting from the retreating upper plate. It is likely that
this domination of compressive strain related to bathy-
metric ridge subduction has dominated over upper platemotion related extension since 15 Ma.
Subduction of the Wharton Ridge initiated at
∼70 Ma ([16], (see Fig. 2(ii)) and has migrated
∼2400 km (30 km/Ma) along the Sunda-Java trench
to its present-day location. During the period 50–15 Ma,
the upper plate adjacent to the location of Wharton
Ridge subduction is both observed and predicted (using
upper plate motions) to have experienced extension.
Plate motions were generally stronger at this time
compared with those affecting Sumatra over the past
15 Myr and so dominated over the compressional effects
of the subducting Wharton Ridge.
The Roo Rise (Fig. 1) is presently being subducting
adjacent to Java. Subduction of this major bathymetric
feature is currently causing deforming the Javanese fore-
arc [49]. Roo Rise subduction is likely to be contributing
to Javanese compression in addition to compression
caused by upper plate advance since ∼15 Ma. Onset of Roo Rise subduction is unknown so the period over
which it has influenced Javanese upper plate strain is
unknown.
3.3. Slab window
A slab window may form between the diverging
plates of a subducting active mid-ocean ridge. Due to
the hotter mantle wedge temperatures expected in
conjunction with a slab window, the viscosity may be
decreased in the mantle wedge and a low viscositymantle wedge can lead to horizontal extension in normal
subduction zones [18,24,50,51]. Our reconstructions
show that a slab window was underlying southern/
central Sumatra at 45–35 Ma (Fig. 2). Our reconstruc-
tions show a minimum slab-window extent as the effects
of slab dip and thermal erosion of plate edges are
excluded, both of which result in increasing the lateral
extent of the slab window. Fig. 2(v–vii) shows the
western edge of the slab window in a stationary position
at the southern tip of Sumatra from 60–45 Ma. It is
likely that thermal erosion of the Indian plate edge over
this 15 million year period would have progressivelyshifted the slab-window edge across southern Sumatra
earlier than 45–40 Ma shown in Fig. 2(vii–viii). Plate
motion reconstructions from 50 Ma (Fig. 3) show an
advancing upper plate in Sumatra suggesting that a
compressive regime should have existed, however
extension is observed from geological evidence. The
underlying slab window may have enabled extension to
continue from ∼50 Ma until the onset of subduction
hinge rollback at 35 Ma.
An underlying slab window can also lead to cessation
of arc volcanism [18], while the progression of a slab edgeacross a region can change chemical signatures, increase
volume and extend the range of volcanism [18].
Plutonism occurred in Indonesia from 60 Ma but was
restricted to Sumatra (and further inland from the trench)
and ceased at around 50 Ma [30,52]. In general, this
Palaeogene volcanic activity was much more prominent
in south and central Sumatra than northern Sumatra [30]
and it has previously been noted that this pattern of
volcanism may be related to subduction of the Wharton
Ridge [16]. It is possible that this underlying slab window
was responsible for the burst of volcanic activity in south
and central Sumatra from 60 to 50 Ma as thermal erosion
454 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 11/13
caused progression of the Indian plate edge across south-
ern Sumatra. Cessation of volcanism at ∼50 Ma suggests
that the slab window was established beneath southern
Sumatra causing volcanic activity to halt. Subduction
related plutonism, focussed along the Sumatran Fault
Zone from the Mid-Miocene [52], re-established inSumatra at around 30 Ma, when that the slab window
was no longer positioned beneath this region.
The slab window underlies Java at 70–40 Ma (Fig. 2
(iii–viii)). At 60–45 Ma extension occurred across
southern Kalimantan and the Java Sea, as well as some
spreading in the Makassar Straits [12,53,54]. Extension
caused by the presence of the underlying slab window
may have exacerbated Javanese back-arc extension,
already induced by a retreating upper plate, enabling
spreading to occur in the Makassar Strait. It is also
possible that the underlying slab window was responsiblefor inhibited Javanese volcanism prior to ∼42 Ma [55].
3.4. Shallow slab dip angle
Fig. 4 shows our reconstructed shallow slab dips
(SSD) plotted against upper plate strain obtained from the
literature for the Indonesian margin. Computed SSD in
the ‘slightly extensional’ category, that fall outside the
present-day relationship of Lallemand and Heuret [13],
represent subduction of young (b23 Ma) lithosphere.
Mid-ocean ridge subduction was excluded from the stud-
ies of Lallemand and Heuret [13] and Sdrolias and Müller [14]. Fig. 4 shows that, if these values are excluded, our
reconstructed regional SSD fall within the pattern ob-
served for the present day by Lallemand and Heuret [13].
Observed SSD (0–125 km) for the Andaman, Sumatra
and Java regions are 24.6°, 23.9°, and 22.7°, respectively
[56], while our calculated SSD are 29°, 25.5°, and 36°,
respectively. The deviation between observed and
calculated clearly shows that the age of subducting
lithosphere alone cannot be used to reconstruct shallow
slab dips, and that further parameters need to be
incorporated into the calculations, such as horizontal
and vertical mantle flow, and down-going plate motions.Therefore, due to errors in the estimation of SSD from the
age of subducting lithosphere, as well as the broad nature
of the relationship between upper plate strain and SSD it is
not possible to use the age of subducting lithosphere to
predict palaeo-upper plate strain regimes.
4. Conclusions
Upper plate strain expected for Sundaland back-arc
regions fromreconstructed trench-normal plate motions of
the Sundaland core and margin correlate well with knownupper plate strain regimes. The three types of upper plate
motion to affect the Sundaland margin since 80 Ma are:
(1) A consistently advancing upper plate corresponds
to compression in the overriding back-arc, caused
by the collision between the down-going Indian
plate and the advancing Sundaland plate,
(2) An advancing upper plate, where the Sundaland
margin advances more rapidly than the Sundaland
core, correlates with extension in the upper plate e.g.
southern Andaman Sea, Sumatra and Java at 30–
15 Ma, 35–15, and 45–15 Ma, respectively. Theonly mechanism for the margin to advance faster
than the core is pulling by subduction hinge rollback,
(3) Uniform upper plate retreat correlates with
extension in the upper plate in two cases, Javanese
crustal extension 60–50 Ma, and spreading in the
Andaman Sea 15–0 Ma.
Fig. 4. This figure shows our shallow slab dip angles plotted against back-arc tectonic regime [11,12,28–31]. The black lines indicate the envelope of
the present-day relationship for worldwide subduction zones from Lallemand et al. [13]. Light grey diamonds represent shallow slab dips for whichthe age of subducting lithosphere is b23 Ma. Dark grey circles are all other data points.
455 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 12/13
Periods where predicted upper plate strain from upper
plate motions does not match observed upper plate strain
can be explained by forces arising from the slab window
and subduction of large bathymetric ridges. In present-
day Java, compression is enhanced by subduction of the
Roo Rise, while in Sumatra, 15–0 Ma, compression fromWharton Ridge and IFZ subduction overrides induced
extension forces from upper plate retreat. The underlying
slab window, formed through subduction of the Wharton
Ridge, may have exacerbated extension in the Java Sea
and south Kalimantan region and possible seafloor
spreading in the Makassar Strait at ∼60–45 Ma. The slab
window may also have enabled extension to continue in
Sumatra from ∼50 Ma until the onset of subduction
hinge rollback at 35 Ma when a uniformly advancing
upper plate could have otherwise led to a compressive
regime. The progression of the slab window across Javaand southern Sumatra also appears to have some
correlation with Indonesian volcanic activity with the
presence of the underlying slab window corresponding
with an absence of Javanese volcanism until ∼42 Ma
[55], and the progressing edge of the slab window
corresponding with an episode of volcanism from 60–
50 Ma in southern Sumatra.
The relationship between our reconstructed shallow
slab dips and known Sundaland upper plate strain
regimes falls within the envelope of observed slab dips
for the present day. However, due to the errors
associated with calculating shallow slab dips from ageof subducting lithosphere, and the broad nature of the
relationship between present-day dips and upper plate
strain, our reconstructed shallow slab dips are not useful
for predicting palaeo-upper plate strain regimes.
References
[1] C.G. Chase, Extension behind island arcs and motions relative to
hot spots, J. Geophys. Res. 83 (1978) 5380–5387.
[2] C.H. Scholz, J. Campos, On the mechanism of seismic
decoupling and back arc spreading at subduction zones,
J. Geophys. Res. 100 (1995) 22,103–22,122.[3] E. Mandovani, D. Albarello, D. Babbucci, C. Tamburelli, M.
Viti, Genetic mechanism of back-arc opening: insights from the
Mediterranean deformation pattern, in: E. Boschi, E. Ekstrom, A.
Morelli (Eds.), Problems in Geophysics for the New Millennium,
Editrice Compositori, Bologna, 2000, pp. 611–643.
[4] M.V. Rodkin, A.G. Rodnikov, Origin and structure of back-arc
basins: new data and model discussion, Phys. Earth Planet. Inter.
93 (1996) 123–131.
[5] D.E. Karig, Origin and development of marginal basins in the
western pacific, J. Geophys. Res. 76 (1971) 2542.
[6] N.H. Sleep, M.N. Toksoz, Evolution of marginal seas, Nature
233 (1971) 548–550.
[7] W.M. Elsasser, Seafloor spreading as thermal convection,
J. Geophys. Res. 76 (1971) 1101–1112.
[8] P. Molnar, T. Atwater, Interarc spreading and Cordilleran
tectonics as alternatives related to the age of subducted oceanic
lithosphere, Earth Planet. Sci. Lett. 41 (1978) 330–340.
[9] J.F. Dewey, Episodicity, sequence, and style at convergent plate
boundaries, in: D.W. Strangway (Ed.), The Continental Crust and
Its Mineral Deposits Special Paper, vol. 20, The Geological
Association of Canada, Toronto, 1980, pp. 553–573.[10] T. Volti, A. Gorbatov, H. Shiobara, H. Sugioka, K. Mochizuki, Y.
Kaneda, Shear-wave splitting in the Mariana trough—a relation
between back-arc spreading and mantle flow? Earth Planet. Sci.
Lett. 244 (2006) 566–575.
[11] J.R. Curray, D.G. Moore, L.A. Lawver, F.J. Emmel, R.W. Raitt,
M. Henry, R. Kieckhefer, Tectonics of the Andaman Sea and
Burma, Geological and Geophysical Investigations Of Continen-
tal Margins, vol. 29, AAPG Memoir, 1978, pp. 189–198.
[12] J. Letouzey, P. Werner, A. Marty, Fault reactivation and structural
inversion. Backarc and intraplate compressive deformations.
Example of the eastern Sunda shelf (Indonesia), Tectonophysics
183 (1990) 341–362.
[13] S. Lallemand, A. Heuret, On the relationships between slab dip,
back-arc stress, upper plate absolute motion, and crustal naturein subduction zones, Geochem., Geophys., Geosyst. 6 (2005),
doi:10.1029/2005GC000917.
[14] M. Sdrolias, R.D. Müller, The controls on back-arc basin
formation, Geochem., Geophys., Geosyst. 7 (2006) Q04016,
doi:10.1029/02005GC001090.
[15] R.D. Jarrard, Relations among subduction parameters, Rev.
Geophys. 24 (1986) 217–284.
[16] C. Heine, R.D. Müller, C. Gaina, Reconstructing the lost Eastern
Tethys Ocean Basin: constraints for the convergence history of
the SE Asian margin and marine gateways, in: P. Clift, D. Hayes,
W. Kuhnt, P. Wang (Eds.), Continent-Ocean Interactions in
Southeast Asia, Geophys. Monogr. Ser., vol. 1149, American
Geophysical Union, Washington, 2004.
[17] K.S. Krishna, D. Gopala Rao, M.V. Ramana, V. Subrahmanyam,
K. Sarma, A.I. Pilipenko, V.S. Shcherbakov, I.V. Radhakrishna
Murthy, Tectonic model for the evolution of oceanic crust in the
northeastern Indian Ocean from the Late Cretaceous to the early
Tertiary, J. Geophys. Res. 100 (1995) 20011–20024.
[18] D.J. Thorkelson, Subduction of diverging plates and the
principles of slab window formation, Tectonophysics 255
(1996) 47–63.
[19] W.R. Dickinson, W.S. Snyder, Geometry of subducted slabs
related to San Andreas transform, J. Geol. 87 (1979) 609–627.
[20] T. Atwater, J. Severinghaus, Relative plate motions in Northeast
Pacific, 85 Ma to present, as recorded by magnetic lineation
patterns, in: Anonymous (Ed.), 28th International Geological
Congress. International Geological Congress, Abstracts—Con-gres Geologique Internationale, Resumes, vol. 28; 1, Interna-
tional Geological Congress, 1989, pp. 60–61, [location varies],
International.
[21] R.G. Bohannon, T. Parsons, Tectonic implications of post-30 Ma
Pacific and north American relative plate motions, Geol. Soc.
Amer. Bull. 107 (1995) 937–959.
[22] D.J. Thorkelson, R.P. Taylor, Cordilleran slab windows, Geology
17 (1989) 833–836.
[23] J.K. Madsen, D.J. Thorkelson, R.M. Friedman, D.D. Marshall,
Cenozoic to recent plate configurations in the pacific basin: ridge
subduction and slab window magmatism in western North
America, Geosphere 2 (2006) 11–34.
[24] V.B. Sisson, A.R. Poole, N.R. Harris, H.C. Burner, T.L. Pavlis, P.
Copeland, R.A. Donelick, W.C. McLelland, Geochemical and
456 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457
7/28/2019 Sunda Java Trench Kinematics
http://slidepdf.com/reader/full/sunda-java-trench-kinematics 13/13
geochronologic constraints for genesis of a tonalite–trondhjemite
suite and associated mafic intrusive rocks in the eastern Chugach
Mountains, Alaska; a record of ridge-transform subduction,
Spec. Pap. — Geol. Soc. Am. 371 (2003) 293–326.
[25] S.T. Johnston, D.J. Thorkelson, Cocos– Nazca slab window beneath
Central America, Earth Planet. Sci. Lett. 146 (1997) 465–474.
[26] V. Ramos, Seismic ridge subduction and topography: forelanddeformation in the Patagonian Andes, Tectonophysics 399
(2005) 73–86.
[27] C. Guivel, Y. Lagabrielle, J. Bourgois, H. Martin, N. Arnaud, S.
Fourcade, J. Cotten, R. Maury, Very shallow melting of oceanic
crust during spreading ridge subduction: origin of near-trench
Quaternary volcanism at the Chile Triple Junction, J. Geophys.
Res. 108 (2003), doi:10.1029/2002JB002119.
[28] C.K. Morley, Combined escape tectonics and subduction roll-
back arc extension: a model for the evolution of Tertiary rift
basins in Thailand, Malaysia and Laos, J. Geol. Soc. (Lond.) 158
(2001) 461–474.
[29] M.G. Bishop, South Sumatra Basin Province, Indonesia: the
Lahat/Talang Akar –Cenozoic total petroleum system, Open-File
Report 99–50 S, U. S. Department of the Interior U.S. GeologicalSurvey, 2000.
[30] R. Hall, Cenozoic geological and plate tectonic evolution of
Southeast Asia and the SW Pacific: computer-based reconstruc-
tions, models and animations, J. Asian Earth Sci. 20 (2002)
353–431.
[31] T. Eguchi, S. Uyeda, T. Maki, Seismotectonics and tectonic
history of the Andaman Sea, Tectonophysics 57 (1979) 35–51.
[32] C. Gaina, R.D. Müller, Cenozoic tectonic and depth/age
evolution of the Indonesian gateway and associated back-arc
basins, Earth-Sci. Rev.(submitted for publication).
[33] C. O'Neill, R.D. Müller, B. Steinberger, On the uncertainties in
hotspot reconstructions, and the significance of moving hotspot
reference frames, Geochem., Geophys., Geosyst. 6 (2005) 002005,
doi:10.1029/2004GC000784.
[34] A.V. Abrajevitch, J.R. Ali, J.C. Aitchison, A.M. Badengzhu, Davis,
J. Liu, S.V. Ziabrev, Neotethys and the India–Asia collision:
insights from a palaeomagnetic study of the Dazhuqu ophiolite,
southern Tibet, Earth Planet. Sci. Lett. 233 (2005) 87–102.
[35] T.-Y. Lee, L.A. Lawver, Cenozoic plate reconstruction of
Southeast Asia, Tectonophysics 251 (1995) 85–138.
[36] G.D. Acton, Apparent polar wander of India since the Cretaceous
with implications for regional tectonics and true polar wander, in:
T. Radhakrishna, J.D.A. Piper (Eds.), Indian Subcontinent and
Gondwana: A Palaeomagnetic and Rock Magnetic Perspective
44, Mem. Geol. Soc. India, 1999, pp. 129–175.
[37] S. Widiyantoro, R. van der Hilst, Structure and evolution of
lithospheric slab beneath the Sunda Arc, Indonesia, Science 271(1996) 1566–1570.
[38] A. Replumaz, H. Karason, R.D.v.d. Hilst, J. Besse, P. Tapponnier,
4-D evolution of SE Asia's mantle from geological reconstruc-
tions and seismic tomography, Earth Planet. Sci. Lett. 221 (2004)
103–115.
[39] R.G. Gordon, C. Demets, J.Y. Royer, Evidence for long-term
diffuse deformation of the lithosphere of the equatorial Indian
Ocean, Nature 395 (1998) 370–374.
[40] P. Patriat, J. Achache, India–Eurasia collision chronology has
implications for crustal shortening and driving mechanisms of
plates, Nature 311 (1984) 615–621.
[41] P. Patriat, J. Segoufin, Reconstruction of the Central Indian
Ocean, Tectonophysics 155 (1988) 211–234.
[42] T.W. Gardner, D. Verdonck, N. Pinter, R.L. Slingerland, K.P.
Furlong, T.F. Bullard, S.G. Wells, Quaternary uplift astride the
aseismic Cocos Ridge, pacific coast of Costa Rica, Geol. Soc.
Amer. Bull. 104 (1992) 219–232.
[43] W.-Y. Chung, H. Kanamori, A mechanical model for platedeformation associated with aseismic ridge subduction in the
New Hebrides arc, Tectonophysics 50 (1978) 29–40.
[44] G.W. Michel, Y. Yue Qui, Z. Sheng Yuan, C. Reigber, M. Becker,
E. Reinhart, W. Simons, B. Ambrosius, C. Vigny, N. Chamot-
Rooke, X. Le Pichon, P. Morgan, S. Matheussen, Crustal motion
and block behaviour in SE-Asia from GPS measurements, Earth
Planet. Sci. Lett. 187 (2001) 239–244.
[45] M.R. Krishna, T.D. Sanu, Shallow seismicity, stress distribution
and crustal deformation pattern in the Andaman–West Sunda arc
and Andaman Sea, northeastern Indian Ocean, J. Seismol. 6
(2002) 25–41.
[46] M. Diament, H. Harjono, K. Karta, C. Deplus, D. Dahrin, M.T.
Zen, M. Gerard, O. Lassal, A. Martin, J. Malod, The Mentawai
fault zone off Sumatra: a new key for the geodynamics of Western Indonesia, Geology 20 (1992) 259–262.
[47] R. McCaffrey, Slip partitioning at convergent plate boundaries
of SE Asia, in: R. Hall, D. Blundell (Eds.), Geological So-
ciety Special Publication, Tectonic Evolution of Southeast
Asia, vol. 106, 1996, pp. 3–18.
[48] L. Prawirodirdjo, Y. Bock, R. McCaffrey, J.F. Genrich, E. Calais,
C. Stevens, S.S.O. Puntodewo, C. Subarya, J. Rais, P. Zwick,
Fauzi, geodetic observations of interseismic strain segmentation
at the Sumatra zone, Geophys. Res. Lett. 24 (1997) 2601–2604.
[49] H. Kopp, E.R. Flueh, C.J. Petersen, W. Weinrebe, A. Wittwer,
Meramex Scientists, The java margin revisited: evidence for
subduction erosion off java, Earth Planet. Sci. Lett. 242 (1-2)
(2006) 130–142.
[50] M.I. Billen, M. Gurnis, A low viscosity wedge in subduction
zones, Earth Planet. Sci. Lett. 193 (2001) 227–236.
[51] D.J. Thorkelson, K. Breitsprecher, Partial melting of slab
window margins: genesis of adakitic and non-adakitic magmas,
Lithos 79 (2005) 25–41.
[52] W.J. McCourt, M.J. Crow, E.J. Cobbing, T. Amin, Mesozoic and
Cenozoic plutonic evolution of SE Asia: evidence from Sumatra,
Indonesia, in: R. Hall, D. Blundell (Eds.), Tectonic Evolution of
Southeast Asia, Geol. Soc. Spec. Publ., vol. 106, Geological
Society of London, 1996, pp. 321–335.
[53] R. Hall, Reconstructing Cenozoic SE Asia, in: R. Hall, D.J.
Blundell (Eds.), Tectonic Evolution Of Southeast Asia, Geolog-
ical Society Special Publications, The Geological Society
London, London, 1996, pp. 153–184.[54] M. Fuller, J.R. Ali, S.J. Moss, G.M. Frost, B. Richter, A. Mahfi,
Paleomagnetism of Borneo, J. Asian Earth Sci. 17 (1999) 3–24.
[55] H. Smyth, R. Hall, J. Hamilton, P. Kinny, East Java: Cenozoic
basins, volcanoes and ancient basement, Indonesian Petroleum
Association Thirtieth Annual Convention and Exhibition, 2005.
[56] C. Cruciani, E. Carminati, C. Doglioni, Slab dip vs. lithospheric
age: no direct function, Earth Planet. Sci. Lett. 238 (2005)
298–310.
457 J.M. Whittaker et al. / Earth and Planetary Science Letters 255 (2007) 445 – 457